05 August 2015

Minnesota: possible first crack in the "Blue Wall"?

Fresh on the heels of a SUSA poll in Kentucky, released yesterday, showing things to be uncomfortably close between presumed DEM nominee Hillary Clinton and favorite son Rand Paul, PPP (D) just released a poll from Minnesota with numbers that, for the first time, really suprised me. I suppose that my Republican friends are going to like to read this information.


GE polling: 1,015 RV, MoE= +/-3.1, of that:
GOP nomination polling: 353 RRV, MoE= +/-5.2 and
DEM nomination polling: 426 RRV, MoE= +/-4.8

GOP nomination:

Walker 19
Trump 18
Bush, Jeb 15
Carson 11
Cruz 7
Huckabee 6
Rubio, Paul 5 apiece
Christie 4
Kasich, Fiorina 3 apiece
Gilmore, Graham, Pataki, Perry and Santorum - all 0 apiece

Margin: Walker +1 over Trump, +4 over Bush, Jeb

With a MoE of +/-5.2, that makes this most definitely a three-way tie between Walker, Trump and Bush. It is, however, the first state nomination polling for the GOP to show someone other than Trump to be in the lead, so I assume that the "favorite son" effect comes into play here.

DEM nomination
Clinton 50
Sanders 32
O'Malley 4
Chafee 3
Webb 2

Margin: Clinton +18

Were this a national poll, people would be whooping and hollering, because a +18 is a landslide exactly on the scale of Ronald Reagan's 1984 win (+18.22%, to be exact), but we have been used to seeing Hillary at +40 against any possible competitors in the DEM field, so by comparison, this looks weak. PPP (D) notes that this is one of Sander's two best polling results anywhere to date, next to New Hamphire, and Sanders tends to do well in the White vote and badly in the minority vote. Both Minnesota and New Hampshire are demographically very white states.

Hillary's fav/unfav in this state is especially bad: 38 fav / 55 unfav = fav -18.
Compared to Trump: 30 fav / 59 unfave = fav -29. His fav/unfav is even far worse.

No candidate wants to be in such a position, but in spite of that, Hillary is ahead of the GOP field in 11 real matchups. That being said, the margins are so far leaner than expected that it is worth reporting.

2016 General Election matchups:

Clinton (D) 43 (49) / Paul (R) 42 (38), margin = Clinton +1 (+11)
Clinton (D) 44 (49) / Bush (R) 42 (39), margin = Clinton +2 (+10)
Clinton (D) 44 (50) / Huckabee (R) 42 (40), margin = Clinton +2 (+10)
Clinton (D) 42 Rubio (R) 40, margin = Clinton +2
Clinton (D) 46 Walker (R) 42, margin = Clinton +4
Clinton (D) 44 Trump (R) 39, margin = Clinton +5
Clinton (D) 44 Cruz (R) 39, margin = Clinton +5
Clinton (D) 44 Carson (R) 39, margin = Clinton +5
Clinton (D) 43 (47) / Christie (R) 38 (37), margin = Clinton +5 (+10)
Clinton (D) 44 / Fiorina (R) 33, margin = Clinton +11
Clinton (D) 41 / Bush (R) 26 / Trump (I) 25, margin = Clinton +15

Clinton wins all 11 matchups, but the first four (vs. Paul, Bush, Huckabee and Rubio) are statistical ties. In matchups against Bush, Bush, Huckabee and Christie, we see a massive reduction in margin over the results of the last PPP (D) poll, but it was released more than one year ago, on 16 June, 2014. Not only that, in a state that has been a bedrock DEM state for 13 of the last 14 presidential cycles, Clinton is nowhere close to the 50 mark. Walker, who until now had not been polled against Hillary in the North Star State, is 4 points behind her, a deficit that a candidate can definitely overcome. Four of those margins are Clinton +5 and only against Fiorina does Hillary have a double digit margin, but look how low the values are. That means there are way too many undecideds in that matchup.

However, a hypothetical three-way matchup against Bush and Trump (as an Independent) looks almost exactly like the 4 or so national three-way matchups we have seen.

Part of the reason why Hilllary is struggling is the fact that she is, for the most part, not doing as well in the womens' vote than Obama did in 2012. Here are the internals from this PPP poll for the women's vote:

PPP Minnesota internals 1.pngPPP Minnesota internals 2.png

Hillary is at +14 above Bush in the womens' vote. You say, "wow, that looks good". But Obama won the women's vote by +19 in 2012 in the final PPP (D) poll and there were, of couse, far fewer undecideds:

PPP Minnesota internals  from final poll 2012.png

But if you also look, Hillary is getting pummeled in the Male vote in this state, whereas Obama, in PPP (D)'s final poll, only lost the male vote by 3 points.

In the headline to the PPP report it says:

"Walker, Clinton Lead in Minnesota; General Closer than 2008 and 2012"

And this is essentially true.

Let's compare these results to polling from 2012, 2008 and 2004, for historical context:


Google Sheets - create and edit spreadsheets online for free.

There were 22 polls of Minnesota in 2012, Obama won all of them. However, in the final polling, one poll, from a Republican firm, showed Obama at only +1:

Minnesota 2012 polling screenshot.png

Please notice that the final PPP (D) and St. Cloud University polls were right on the mark. But you will notice other polls showing quite lean margins for the President in 2012, like the Star Tribune and Rasmussen. Obama won the state by +7.7 points, which rounds to +8; both PPP and St. Cloud nailed both toplines AND the margin. So, when PPP (D) shows the race in Minnesota currently to be a real dogfight, I believe it.


Statistikhengst s ELECTORAL POLITICS - 2015 and beyond FINAL POLL CONVERGENCE No. 12

In 2008, there were 51 polls taken in Minnesota and Obama won 47 of them. He won the state by +10.24 points in that year, but the end polling was all over the place:

Minnesota 2008 polling screenshot.png

The end-polling aggregate showed Obama +12, so the aggregate was off to the Left by +1.76 points, not great, but also not terrible. In that year, the Tribune poll came closest to the actual results and SUSA, which showed +3, was way off. PPP (D), which was new to the scene that year, was also off to the Left by 5 points, which is bad. This is a factor in polling that PPP (D) obviously corrected between 2008 and 2012.


RealClear Politics - Polls

Minnesota 2004 polling screenshot.png

In 2004, Kerry won Minnesota by +3.48%, so it was damned close to the aggregate provided by RCP in that year, which showed Kerry +3.2. This is once again proof of why I trust aggregates more than others do, because in spite of some wild polling in 2004 that was way off either one way or the other, the average came very close to the truth, as it did in 2008 and definitely in 2012. The aggregate in 2012 was only 1 point to the Right, which is completely acceptable.

So, were Clinton vs. GOP field in this state to continue to look like this one PPP (D), I would tend to say that the race here would be looking more like 2004 than either 2008 or 2012, which means that if these figures continue to be so, Minnesota would become a battleground state.

That being said, even in 2008, a year where Obama landslided here, you can find a poll showing only +1, and in 2012, a poll showing only +3, which is why we need to see more polling from this state.

Historical background:

The electoral history of Minnesota since 1960 is also very enlightening to see:

Electoral results Minnesota 1960-2012.png

That is a sea of blue in this state. Since 1976, 10 cycles in a row, this state has cast its electors for the Democratic presidential nominee and since 1960, 14 cycles, it is a 13-for-14 D-state. It is the only state in 1984 (outside of DC) that Reagan didn't carry and it took Nixon a +23 national win in order to pull Minnesota over the line, by +5.51%. It was also his leanest win of that cycle. So, the standard wisdom is that if a Dem is really, really struggling in Minnesota, he is probably losing nationally. See 1980, 1984, 2000 and 2004 as examples. But that's also not totally true because in 1988, Michael Dukakis, a "Massachusetts Liberal", easily won Minnesota but yet, lost the 1988 GE. So, there is no hard and fast formula for this state. Specifically, Dukakis won Minnesota by 7 points in 1988 but lost to Bush by almost 8 points in that year, meaning that Minnesota went about 15 points to the Left of the national results in that year. But in 2012, Obama won MN by almost 8 points and won the General by 4, so Minnesota was only 3 points to the left, just as it was in 2008 and just as it was in 2000. So, the state is not as Liberal leaning as it was in 1996, to be sure.

Could the GOP win Minnesota in 2016? Sure it could, anything is possible. In many ways, Minnesota is to the Democrats as Indiana is to the Republicans. And to prove that point, here is the electoral history of Indiana, also a state from the Midwest, also a state similar in population to Minnesota:

Electoral results Indiana 1960-2012.png

Indiana is almost a reverse image of Minnesota. Here, we see a sea of red. From 1968-2004, Indiana was a 10 for 10 GOP just as Minnesota now is a 10 for 10 DEM state, and overall, from 1960-2012, it's a 12-for-14 GOP state. You will notice that the margins in Indiana, excepting Johnson's 1964 national landslide, tend to be much bigger for the GOP in this state than the DEM margins in Minnesota.

Now, let's put the two states side-by side:

Electoral results Indiana vs Minnesota 1960-2012.png

The contrast is pretty darned stark, eh? Since 1960, 14 presidential cyles, only 3 times have both parties gone for the same candidate, and in all three cases, that candidate won the national election: 1964, 1972 and 2008. In two of those cases (1964, 1972) the case of Johnson sweeping Republican Indiana and Nixon sweeping Democratic Minnesota was a sure sign of a massive national landslide. In 2008, it's was different. Obama's statistic in Republican Indiana was somewhat amazing in that the state shifted 21.71 points to get from Bush 43's 2004 statistic to Obama's 2008 statistic, and yet, his +7.26% national win was, in terms of the NPV, not a landslide.

Now, look at the field that is called "Diff." (difference). That is the margin between these two states. In the case of like colors, you subtract. In the case of different colors, you add. For instance, in 1968, Humprey won MN by+12.53 points. You can consider that as 12.53 points to the LEFT of the "50 yard line". Nixon won IN at the same time by +12.30 points (almost a perfect mirror image of MN), and you can consider his margin to be 12.30 points to the RIGHT of the "50 yard line". Difference = 12.53 + 12.30 = 24.83. You will see that, apart from 2008, the difference between this two states is always in double-digits and virtually all the time, above 17 points. So, in terms of voter behavior, these states are about as different as they come, but share one quality: their reliability for their respective "team".

This brings me to the key point I am wanting to make:

in 2008, when polling showed Obama at par or ahead of McCain in INDIANA, that was probably the surest sign of all that McCain was heading for a resounding loss.

And the same applies to Hillary Clinton. Should state polling start to show her really losing in MINNESOTA, then the probablility that she is losing nationally is pretty much 100%.

Now, I just spent a lot of time over just one poll, but this is important because for the first time, we see polling data from a core DEM state that shows a possible, real crack in the so-called "Blue Wall" that has existed since 1992.

For this reason, I will be keeping a very watchful eye on Minnesota.

No comments:

Post a Comment

Constructive comments and critique are always welcome. Please keep it polite and respectful.